opennlp.tools.ml.model.OnePassDataIndexer Maven / Gradle / Ivy

Go to download
Show more of this group Show more artifacts with this name
Show all versions of opennlp-tools Show documentation
The newest version!
/*
 * Licensed to the Apache Software Foundation (ASF) under one or more
 * contributor license agreements.  See the NOTICE file distributed with
 * this work for additional information regarding copyright ownership.
 * The ASF licenses this file to You under the Apache License, Version 2.0
 * (the "License"); you may not use this file except in compliance with
 * the License. You may obtain a copy of the License at
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */

package opennlp.tools.ml.model;

import java.io.IOException;
import java.util.HashMap;
import java.util.LinkedList;
import java.util.List;
import java.util.Map;

import org.slf4j.Logger;
import org.slf4j.LoggerFactory;

import opennlp.tools.util.ObjectStream;
import opennlp.tools.util.ObjectStreamUtils;
import opennlp.tools.util.TrainingParameters;

/**
 * A {@link DataIndexer} for maxent model data which handles cutoffs for uncommon
 * contextual predicates and provides a unique integer index for each of the
 * predicates.
 *
 * @see DataIndexer
 * @see AbstractDataIndexer
 */
public class OnePassDataIndexer extends AbstractDataIndexer {

  private static final Logger logger = LoggerFactory.getLogger(OnePassDataIndexer.class);

  public OnePassDataIndexer() {}

  /**
   * {@inheritDoc}
   */
  @Override
  public void index(ObjectStream eventStream) throws IOException {
    int cutoff = trainingParameters.getIntParameter(TrainingParameters.CUTOFF_PARAM,
        TrainingParameters.CUTOFF_DEFAULT_VALUE);
    boolean sort = trainingParameters.getBooleanParameter(SORT_PARAM, SORT_DEFAULT);

    long start = System.currentTimeMillis();

    logger.info("Indexing events with OnePass using cutoff of {}", cutoff);

    logger.info("Computing event counts...");
    Map predicateIndex = new HashMap<>();
    List events = computeEventCounts(eventStream, predicateIndex, cutoff);
    logger.info("done. {} events", events.size());

    logger.info("Indexing...  ");
    List eventsToCompare =
        index(ObjectStreamUtils.createObjectStream(events), predicateIndex);

    logger.info("done.");

    logger.info("Sorting and merging events... ");
    sortAndMerge(eventsToCompare, sort);
    logger.info(String.format("Done indexing in %.2f s.", (System.currentTimeMillis() - start) / 1000d));
  }

  /**
   * Reads events from eventStream into a linked list. The predicates
   * associated with each event are counted and any which occur at least
   * cutoff times are added to the predicatesInOut map along
   * with a unique integer index.
   *
   * @param eventStream
   *          an EventStream value
   * @param predicatesInOut
   *          a TObjectIntHashMap value
   * @param cutoff
   *          an int value
   * @return a TLinkedList value
   */
  private List computeEventCounts(ObjectStream eventStream,
      Map predicatesInOut, int cutoff) throws IOException {

    Map counter = new HashMap<>();
    List events = new LinkedList<>();
    Event ev;
    while ((ev = eventStream.read()) != null) {
      events.add(ev);
      update(ev.getContext(), counter);
    }

    String[] predicateSet = counter.entrySet().stream()
        .filter(entry -> entry.getValue() >= cutoff)
        .map(Map.Entry::getKey).sorted()
        .toArray(String[]::new);

    predCounts = new int[predicateSet.length];
    for (int i = 0; i < predicateSet.length; i++) {
      predCounts[i] = counter.get(predicateSet[i]);
      predicatesInOut.put(predicateSet[i], i);
    }

    return events;
  }
}